Notes On Manuscript Transmission Of Vaisesika Sutra And Its Earliest Commentaries

Author(s): Harunaga Issacson
Publisher: Harunaga Issacson
Catalog link: https://jainqq.org/explore/269539/1

Page #1
NOTES ON THE MANUSCRIPT TRANSMISSION OF THE VAISESIKASUTRA AND ITS EARLIEST COMMENTARIES Harunaga Isaacson (University of Groningen) In view of the textual problems commentary on the Vaisesikasurrounding the Vaisesikasutra an sutra. It is shown that a reexamination of the surviving ma- examination of the manuscripts nuscript evidence is an urgent can lead to improvement of the desideratum, as was emphasized text. Two manuscripts not used by A. Wezler in an article pub- in the published edition are introlished in 1982. A start in this di- duced. Section V contains obserrection has been made, and some vations on the two versions of the of the findings thus far are pre- commentary by Bhatta Vadindra sented in this paper. Section I in- on the Vaisesikasutra. Substantroduces the problem and summa- tial improvements over the pubrizes earlier work. In sections II lished text of the abridged version and ill two manuscripts contain- proved to be possible, especially ing the sutrapatha alone are re- with the aid of the palm-leaf maported on; both are shown to di- nuscript which was not available verge extensively from the pub- to the editor. Section vi concludes lished recensions of the text. Sec- by offering some general remarks, tion IV discusses Candrananda's chiefly on questions of method. Anyone who attempts to study the Vaisesikasutra (VS) will soon enough be confronted with problems resulting from the defective transmission of this text. This is a fact that has often been remarked on, but to date the best summary of the situation is that found in the opening pages of A. Wezler's article in the Festschrift for J.W. de Jong. Among other points, Wezler For access to or copies of manuscripts referred to in this article I am indebted to the authorities of the Bhandarkar Oriental Research Institute (Poona), the L.D. Institute (Ahmedabad), the Oriental Institute, M.S. University of Baroda, the Asiatic Society, Calcutta, and Kerala University Manuscripts Library (Trivandrum). I am very grateful to Prof. Dr. A. Wezler for help in acquiring copies of several manuscripts and for his kind encouragement in the work reported on here. Wezler 1982, 643-648. Among earlier publications which discuss the general problems posed by the textual situation of the VS, one might mention in particular Thakur 1963a and the introduction contributed by Thakur to Muni Jambuvijaya's edition of the VS together with Candrananda's commentary.

Page #2
emphasized the importance of examining manuscripts containing the text of the VS alone, i.e. without a commentary, in view of the possibility that some such manuscripts might either represent an independent transmission of the sutrapatha or contain a text which was extracted (uddhrta from a commentary which preserved at least a better text than that of the 'Maithila version,' i.e. that commented on by Sankara Misra. Wezler concluded his observations on the manuscript transmission of the VS with the following paragraph. Since I do not intend, or rather am not able at present, to carry out this indispensable examination of all the MSS of the VS, I shall not dwell on this point any longer. This much only I should like to add by way of summary: the transmission of the VS has unfortunately been of such a kind that even the faintest opportunity should not be missed to enlarge the documentary basis on which a critical edition of this important text ought to be built. Though well known, the fact bears repetition: elementary philological work done till now in the field of Indian philosophy is quite inadequate and unsatisfactory. (Wezler 1982, 645) These words were written and published already more than a decade agoyet, as far as I am aware, the 'indispensable examination of all the MSS of the VS' has still not been carried out. Nor have I been able to do so, but for some time now I have been making efforts to examine as many manuscripts of the VS and its commentaries, as well as of Prasastapada's Padarthadharmasargraha, as I could gain access to, either directly or in the form of copies of some kind. Though this study is still far from being complete and exhaustive, the manuscripts thus far examined include a number which do indeed seem to provide new and significant evidence for the text of the VS, so that I believe it may be worthwhile to present a preliminary report. As to manuscripts containing the text of the VS without an accompanying commentary, the only two scholars who have published information so far, to the best of my knowledge, are Gopinath Kaviraj and Anantalal Thakur.3 In a brief article which was published as long ago as 1929, but which has been perhaps somewhat undeservedly neglected," Gopinath Kavi 2 The manuscript tradition of this text forms a separate problem, and one of a different nature. A discussion must be postponed till a future occasion. * Their relevant publications are also referred to by Wezler in the article which was quoted above: see p. 643-644, note 3 on p. 674, and n. 8 on p. 675. *Nozawa's Comparative Table of the Vaisesikasutra (1985), for instance, does not report

Page #3
raj reports on 'an apparently very old manuscript (undated) of the Vaisesika Sutras,' from a private collection in Benares. Kaviraj noticed that the text of this manuscript contained 'several differences from the current text' (i.e. the text as commented on by Sankara Misra) and in his article gives, so he says, the differences. Regrettably, a complete transcript of this manuscript was never published, and its present whereabouts are not known to me-it is not unlikely that it may never be traced. We must therefore rely solely on Kaviraj's collation, the completeness of which is uncertain at best. It is noteworthy that in several cases the reported text contains lacunae, and it seems therefore very likely that also in cases where Kaviraj found no clear difference from Sankara Misra's text to note, the manuscript may have been in fact damaged or illegible. I am therefore very doubtful as to whether for sutras for which Kaviraj does not give a variant reading from the manuscript, we may always safely conclude es silentio that it read as does Sankara Misra. It would appear to be Anantalal Thakur who has gone to the most trouble to examine manuscripts of the VS and its commentaries, as well as to exert himself laudably in the work of editing. According to an article with the title Textual Problems of the Vaisesikasutras, published in 1963, Thakur had at that time 'collated the sutra-readings from six printed editions and sixteen Manuscripts preserved in different Manuscript Libraries. Unfortunately, Thakur does not give details of the manuscripts he collated, not to speak of their readings. The sentence immediately following on the one quoted also has a confusing rather than an enlightening effect; Thakur says that among them two have subsequently formed the basis of the editions of the Vaisesikadarsana published by the Mithila Institute, Darbhanga and the Oriental Institute, Baroda.' The two editions referred to must of course be Thakur's own edition of the abridged version of Bhatta Vadindra's commentary (V) and Muni Jambuvijaya's edition of Candrananda's commentary (C), respectively. But since two manuscripts were used by Jambuvijaya and one by Thakur, these two editions are based on three rather than two manuscripts, as Thakur states here. Furthermore, it appears from this that the sixteen manuscripts mentioned include manuscripts of the VS together with the divergent readings Kaviraj quoted in this article. Kaviraj 1929, 71. The name of the owner of the collection is not given because, as it appears, he wished to remain anonymous. Kaviraj tells us that he was able to use the manuscript for a few days only (p. 71). ?Thakur 1963a, 187. 8 This text and Thakur's edition are discussed in section v below.

Page #4
commentaries, but it is not made completely clear whether also manuscripts containing only the sutrapatha were collated. Nor can it be excluded that the manuscripts Thakur referred to included some of the VS together with Sankara Misra's commentary. Earlier, in the introduction to his edition of V, Thakur had stated that "[t]he known manuscripts of the Vaisesikasutras are not numerous. They generally represent the Maithila version just mentioned." Here too, Thakur unfortunately gives no information as to the exceptions the existence of which he implies, and once more the possibility cannot perhaps be excluded that Thakur had in mind manuscripts giving the text of the VS together with commentaries (for instance those of Candranada-at the time known of but not published-and Bhatta Vadindra) as well as manuscripts of the sutrapatha alone.10 In short, Thakur's publications hardly give us any concrete information as to manuscripts which give a sutrapatha alone and differ from the text followed by Sankara Misra. The hope need not yet be given up that Thakur one day will do so, or even publish the critical edition he had been planning, or else collations of all the manuscripts he has examined, but as the years pass, the chance of this happening becomes ever slimmer. II My examination of manuscripts containing the sutrapatha without a commentary has confirmed Thakur's remark as to the prevalence of the version commented on by Sankara Misra, but two manuscripts I have been able to collate have proved highly interesting exceptions. Both contain texts which differ from the known recensions of the VS, as well as being mutually quite different. The publication of a complete 'edition' of these two MSS is envisaged in the near future; this section and the following one aim at briefly introducing them and demonstrating, by means of quotes, their independence from the known commentaries. The first manuscript I shall deal with is a 'Sammelhandschrift' in the L.D. Institute, Ahmedabad, hereafter designated as A. The first text in the manuscript is that of the Nyayasutra, without a commentary. This is This remark is found on p. 11 of the English introduction. The corresponding passage in the Sanskrit bhumika reads upalabhyamanah sutramatrkast prayaso maithilapathanusarinyah (p. 24). 10In another article we again find the statement that '[o]ld manuscripts of these sutras are rare and those available generally follow the Vaisesikasutropaskara of Sankaramisra (15th cent. A.D.)' (Thakur 1963b, 78). But here too, no details are given after this general statement. 4

Page #5
followed by the VS, again without a commentary. Only the folios containing the Nyayasutra and the VS, together with the beginning of another work which I have not yet identified, are available to me at present, in the form of a photocopy (made from microfilm) kindly provided to me by Prof. A. Wezler. Hence I shall not, indeed can not, here provide a full description of the manuscript. The manuscript is written in Jaina Devanagari script and by a single hand. The portion available to me bears no date. I am skeptical about the possibility of dating it on purely palaeographical grounds, but, for what it's worth, my personal judgement would be that the hand is relatively early; that it is to say, I should be a little surprised if it were to prove to be later than the seventeenth or early eighteenth century. The text of the VS begins on folio 4 and ends on f.7". The individual sutras are not numbered, nor is there always a single or double danda after them. On the other hand, there are occasional dandas in the middle of what must, on considerations of sense as well as in view of the other recensions, be a single sutra. I may remark that this, as well as the fact that sandhi is regularly applied between the end of a sutra and the beginning of the next, suggests that the text in this manuscript probably was not extracted directly from a manuscript containing the sutras embedded in a commentary. For if we assume that the scribe of A went through a manuscript containing both sutras and commentary and copied out the sutras alone, it follows that he would have had to be able to identify the sutras in the examplar he was copying from and recognize where each sutra ended and the commentary began. Therefore the signs I mentioned, suggesting that in fact the scribe does not always identify the ends and beginnings of the sutras correctly, speak against this theory. Of course it remains perfectly possible that an 11 Two other scenarios are at least as unlikely. One might consider the possibility that the scribe had before him a manuscript of a commentary which did not give each sutra separately, followed by its commentary, but merely contained occasional pratikas of the sutras. This I find highly unlikely because I cannot credit that the result of such a scribal reconstruction of the sutras would have been nearly as good as A in fact is. Furthermore, none of the manuscripts I have examined of the VS together with a commentary are in fact of this type. One more possibility could be that the manuscript was dictated to the scribe, whether by someone using a manuscript with sutras and commentary, by someone using a manuscript with the sutrapatha alone or from memory. However A contains enough errors which point to copying from another manuscript for this hypothesis to be quite unconvincing. For instance, we find some clear cases of misreading of similar aksaras, as well as of probable eyeskip.

Page #6
ancestor of A was extracted from a commentary in some way or other. In the following sections, references to sutras use the numbering of C unless otherwise stated. In transcribing from manuscripts I add word divisions but make no further changes or additions. Consonants written with a virama (which may in some cases be an indication of sutra division) have been indicated as such by a line under the letter concerned, e.g. t. The text of the VS followed by Bhatta Vadindra is designated as BhV when based on the long version of the commentary and V when based on the abridged version. These two versions are discussed in section v below. The readings of the VS found in the so-called 'Sena Court' commentary, available on adhyayas nine and ten only, are referred to by the siglum S. 1 A omits 1.1.4 of SM, dharmavisesaprasutad dravyagunakarmasamanya visesasamavayanam padarthanam sadharmyavaidharmyabhyam tattva jnanan nihsreyasam. 2 2.1.26 reads lirigam akasasya in A, agreeing with C and diverging from BhV/V's sabdo lingam akasasyeti and SM's parisesal lingam akasasya. 3 4.1.13 in A reads arupisu acaksusatvat, in agreement with C and V and differing from SM's arupisu acaksusani. 4 In 5.2.21 and 5.2.22 A has dravyagunakarmmanispattih (read nispattio) vaidharmmyad bhasa abhavas tamas tejaso dravyamtarenavaranac ca tamah. In place of this C reads dravyagunakarmavaidharmyad bhavabhavamatram tamah (5.2.21), tejaso dravyantarenavaranac ca (5.2.22). V has merely dravyagunakarmanispattivaidharmyad bhabhavas tamah,13 with no counterpart for the second sutra. SM dravyagunakarmanispattivaidharmyad abhavas tamah and tejaso dravyantarenavaranac ca. 5 7.1.12, which reads in Cagunavato dravyasya gunarambhat karmaguna agunah, and in V karmaguna aguna, is found in A in the following form: agunavato dravyasya gunarambhat karmmana gunah. The sutra has no equivalent in SM. 12 The ninth adhyaya of this commentary has been published as an appendix to Thakur's edition of BhV. The sutrapatha of the tenth adhyaya is given in Thakur 1965; I have checked it against the manuscript. 13 And not vaidharmyad abhavas tamah as reported in Nozawa 1985, 85.

Page #7
6 The eighth adhyaya is divided into two ahnikas, with the sutra artha iti dravyagunakarmasu (= 8.14) the last one in the first ahnika and the sutra dravyesu pancatmakam pratyuktam the first of the second. In this A differs from the other recensions: C does not divide this adhyaya into ahnikas, 14 V starts the second ahnika one sutra earlier, with artha iti gunadravyakarmasu and SM starts it earlier yet, with (C's) 8.12, ayam esa tvaya kttam bhojayainam iti buddhyapeksam. Note by the way that in this sutra, for SM's tvaya kTtam C reads kTtam tvaya and V as well as A simply tvaya. 7 A does not divide the ninth adhyaya into ahnikas. In this it agrees with C and (probably) S;15 V unfortunately is lost after [9.7 = C's] 9.8, but the commentary on this sutra is followed by a colophon of the first ahnika of the adhyaya. SM divides into ahnikas, taking 9.18, asyedam karyam karanam samyogi virodhi samavayi ceti laingikam as the opening sutra of the second ahnika. 8 9.1 in A reads kriyagunavyapadescbhavad asat, as also found in C and V. SM and S read kriyagunavyapadesabhavat pragasat. 9 The tenth adhyaya is again divided into two ahnikas in A. The sutra laingike pramanam vyakhyatam is the final one of the first ahnika; this corresponds to C's 10.19, which has laingikam for laingike. S however reads as does A; V is again not available. The sutra has no counterpart in SM, where the second ahnika begins with C's 10.12, karanam iti dravye karyasamavayat. Note that with the ahnika division found in A. the second ahnika is reduced to a mere two sutras; and, perhaps *And in this, as has often been remarked, agrees with the brief description of the VS given in Madhava's Sarvadarsanasamgraha. 15 One cannot perhaps be completely certain about S, for one folio, folio 31, appears to be lost in the unique manuscript. The last sutra on folio 300 is 9.10; the first on folio 32" is 9.15. It can therefore not be determined which of the intervening sutras were actually known to the commentator (note that C's 9.11 and 9.12 are not in SM's text), nor can it be completely excluded that the missing folio contained a colophon for a first ahnika of 9. The fact that SM divides the adhyaya elsewhere, after (C's) 9.17, does not rule out this possibility, for we already saw with regard to the eighth adhyaya that the recensions which do divide into ahnikas do so at different places. The fact that the final colophon of 9 in S does not mention ahnikas is also inconclusive; the same may be said of most of the adhyaya colophons in the manuscripts of Candrananda's commentary on the VS, even in the adhyayas which do consist of two ahnikas. None the less, I think it likely that S indeed did not accept such a division; note that none is found in the manuscript of the tenth ahnika of S.

Page #8
significantly, these sutras are the two which occur earlier in the VS. The sutra drstanam drstaprayojananam drstabhave prayogo 'bhyudayaya (C's 10.20; A, S and SM all have the same reading too) occurs earlier as 6.2.1, while the final sutra tadvacanad amnayapramanyam (thus C, S16 and A; SM reads tadvacanadamnayasya pramanyam) is VS 1.1.3. 10 10.8 in A reads abhud ity abhutat, agreeing with C and S, and differing from SM which reads abhud ity api. V is not available here. The cases listed above, though a mere sample, 17 should I think be sufficient to establish that A represents a hitherto unknown recension of the VS, and one which is in numerous respects superior at least to the version commented on by Sankara Misra. Among the commentaries, A's text is decidedly closest to that followed by Candrananda, but the differences between the two versions, such as those noted under points 4, 5, 6 and 9 above, are too many to allow us to regard them as following the same recension. INI Another manuscript which contains the text of the VS with no accompanying commentary is a palm-leaf manuscript in the Kerala University Manuscripts Library, Trivandrum.18 I shall refer to this MS in the following as T. As in A, the text of the VS is preceded by that of the Nyayasutras, again without a commentary.19 16 Thakur reports S as reading tadvacanad amnayasya pramanyam with SM (Thakur 1965, 21). But this is incorrect; the manuscript (which I have consulted from photocopies kindly provided by the Asiatic Society, Calcutta) is a little difficult to make out but definitely reads tadvacanad amnayapramanyam. 17A's readings of a number of other sutras are quoted, by way of comparison, in several of the examples given in the following sections below. 18 The manuscript number is 22615B, although the photocopy kindly supplied to me erroneously has the number 921B written on it. It appears to be uncatalogued; it is not listed in the Alphabetical Inder of Sanskrit Manuscripts in the Oriental Research Institute and Manuscripts Library, Trivandrum. Vol. III (Ya to Sa) (Bhaskaran 1984). I am not sure what conclusions, if any, may be drawn from the fact that both A and Tare Sammelhandschriften.' From having stumbled on these two cases in the course of my really rather limited examination of VS manuscripts, I suppose that there may well be other such manuscripts containing, for example, the text of the Nyayasutra and the VS, perhaps together with other texts. Unfortunately, such manuscripts are at a greater risk than most of being wrongly catalogued, since correct identification depends on the manuscript being gone through more carefully than the glance at beginning and end which is often all that a cataloguer will find time to do. 8

Page #9
Once more, I refrain from attempting to give a thorough description of the manuscript, in view especially of the fact that I have access only to photocopies of the folios which contain the text of the VS. The VS covers folios 20"-34". The script is Malayalam. The manuscript bears no date but is in good condition; from its general appearance as well as on the basis (admittedly uncertain) of palaeography I should hazard that it is no older than the nineteenth century. Punctuation marks, usually small dots between the aksaras, are occasionally found, but by no means between all the sutras. Similar considerations as set out in regard to A above lead me to believe that the examplar from which the manuscript was copied also contained the sutrapatha with no commentary. Unfortunately, the number of scribal errors and corruptions in T is far greater than in A, so that in many cases it is not possible to be certain of the intended reading. None the less, the following are some of the interesting readings which feature in this manuscript, which seem to me to justify speaking of yet another recension. 1 T too does not contain SM's 1.1.4, but instead reads a different sutra, found in no other source known to me, after 1.1.3: sadhanany asya dravyagunakarmmani. This sutra, which no doubt should not be regarded as 'original,' seems to serve a purpose somewhat similar to that of SM's 1.1.4. That is to say, its inclusion may be motivated by the desire to have the sutras state their subject matter (abhidheya) at their outset more clearly than is done in 1.1.1 (athato dharmam vyakhyasyamah; thus all recensions, supported by numerous testimonia), as well as to indicate the connection (sambandha) between the subject matter of the VS and the ultimate goal (prayojana), which is understood from 1.1.2 (yato 'bhyudayanihsreyasasiddhih sa dharmah; thus, bar orthographical variants and obvious slips, all recensions, again supported by several testimonia) to be both worldly and supreme good. Frauwallner apparently found it inconceivable that the 'original' text of the VS should fail to name the categories of the Vaisesika; 20 those responsible 20 Frauwallner's keen philological instinct may perhaps have erred for once when he wrote 'In den Vaisesika-Sutren mit dem Kommentar des Candrananda (VSu1) und mit dem anonymen, von Anantalal Thakur veroffentlichten Kommentar (VSu2) [i.e. V] fehlt das vierte Sutram. Doch ist am Anfang des Textes eine Nennung der sechs padarthah unerlasslich' (Frauwallner 1984, 36-37 n. 5). It is precisely the absence of the expected enumeration of categories which is likely to be original here. Indeed an enumeration of siz categories would be suspect, for I think it very likely that in the earliest period of composition of sutras the classical list of padarthas had not yet been settled on.. 9

Page #10
for adding this sutra and SM's 1.1.4 may well have thought much the same.21 It is interesting, however, that the added sutra in T has a perhaps slightly archaic ring to it, in that it enumerates only the first and most fundamental three categories, unlike SM's 1.1.4. 2 The sutras 2.2.4-5 read in C and A tejasy usnata (2.2.4), apsu sitata (2.2.5). For tejasy usnata V reads tejahsusnata and SM tejasa usnata. T is unique in reversing the order of these sutras; its reading is apsu sitata tejasy usnata. 3 2.2.16 is found in T in the following form: adityasamprayogat bhavisya to bhutac ca praci. The edition of C reads adityasamyogad bhutapurvad bhavisyato bhutac ca praci (but see section iv below), as do SM and A, while BhV/V has adityasamyogad bhutapurvad bhavisyato bhutac ca.22 4 T reads 3.1.9 as follows prasiddhabhutapurvakatvad apadesasya. This differs from the other versions; C prasiddhapurvakatvad apadesasya, V, SM and A prasiddhipurvakatvad apadesasya. 5 Between the sutras 6.1.4, buddhipurvo dadatih (thus C, V, SM and A; T, though reading buddhipurvo dadati should also be corrected to read thus) and 6.1.5, tatha parigrahah (thus all versions), T inserts what appears to be a hitherto unknown sutra mahine capravrttih. Here the possibility should be considered that this sutra has arisen due to some form of textual corruption. Compare 6.1.14 same hine capraurttih (thus C and T: A same hino capravsttih, which should be emended to agree with Cand T; SM same hine vapravsttih). If some sort of eyeskip forward and again back is indeed the source of this 'sutra' (though it is hard to explain such an occurence here), this would tend to confirm that the examplar also contained the strapatha only. 6 7.2.14 reads as follows yutasiddhyabhavat karyakaranayos samyogavi 21 And such considerations may well underlie the pratijria attributed to Kanada in the sentence, quoted twice by Vyomasiva in his Vyomavati, with slightly different wordings, which Frauwallner supposed to be the original opening of the VS. Cf. Frauwallner 1984 and Halbfass 1992, 69-70. 22 The suggestion of Nozawa that V should read ...ca na praci (Nozawa 1974, 472 and Nozawa 1985, 79) is ruled out rather than confirmed by the publication of BhV; cf. BhV p. 269. 10

Page #11
bhaga na vidyante. C, V23 and SM all read yutasiddhyabhavat karyakaranayoh samyogavibhagas na vidyete. A reads yutasiddhyabhavat karyakaranayoh samyogavibhago na vidgate, but this should probably be taken to be a scribal error for ... samyogavibhagau na vidyete. 7 In 7.2.16 T agrees with C and V, as well as A in reading gune ca bhasyate. SM reads guno 'pi vibhavyate. 8 7.2.17 is read unanimously by C, V, SM and A as niskriyatvat. In T we find the following: sbhayopagamanan niskriyatvad atitanagatapratyayabhavat prasamgat. We cannot be certain, but this should perhaps be taken as four separate sutras, three of which are not known to me from any other source. If they have been introduced from some commentary, it must be one which has not yet been discovered, for I could find nothing in the commentaries by Candrananda, Bhatta Vadindra or Sankara Misra which even vaguely resembled these sutras. 9 The eighth, ninth and tenth adhyayas are not divided into ahnikas in T. From the above examples it will be gathered that T is an interesting and rather eccentric manuscript. Its differences from the other recensions are usually more radical than those of A. Like A, it contains many features which make an older impression than the text of SM. Despite the fact that the manuscript is not a very correct one, the divergent readings and extra sutras it appears to contain deserve to be taken seriously and judged on their own merits. The possibility that the recension represented by T is an old one cannot be ruled out; as far as we can tell, different versions of the VS were in existence already at an early period. IV In addition to manuscripts containing the text of the VS alone, those containing the sutras together with a commentary should also be collected and examined as thoroughly as possible. It may not be vain to hope that one day a hitherto unknown commentary, that of Atreya for example,24 may 23V should be corrected to read thus, as indicated by Nozawa 1974, 471, and in fact already by Thakur himself in the second appendix (giving the sutrapatha) of his edition. The edition itself, as well as the reprint of the text in the appendix of Thakur's edition of BhV, reads vidyate for vidyete. 24 The best source of information on this commentary to date is formed by the fairly numerous quotations or references to it in the commentary by Bhatta Vadindra. 11

Page #12
yet be found to exist in manuscript form. And besides, we should not neglect to look for material which might allow improvement of the texts of the commentaries already available to us. That such improvement is possible in regard to the texts of our two oldest commentaries, by Candrananda and Bhatta Vadindra respectively, is what I shall try to demonstrate in this section and the following one. The publication in 1961 of the commentary on the VS by Candrananda formed a landmark in studies of the Vaisesika. A recension of the VS was hereby made available that was clearly superior to that represented by Sankara Misra as well as that of Bhatta Vadindra, which had been published a few years earlier. Besides, the commentary too presented us with several interpretations which, in their simplicity, seemed superior to those of the later scholiasts. Finally, the text was fortunate in its editor, the Jaina luni Jambuvijaya, perhaps the most distinguished scholar to edit a Vaisesika text. Small wonder then that this publication was received with gratitude and admiration by the most prominent scholars working in the field. 25 Jambuvijaya's edition is indeed a good one, perhaps one of the most satisfactory editions of a classical Indian philosophical text. Still, it may be truely said that no edition is ever really definitive, and in the course of examining the manuscripts of Candrananda's commentary I have been brought to the conclusion that further progress is possible in regard to this text. At present I am working on a new edition of the commentary. That this is not wholly superfluous labour I shall try to demonstrate in the following. But if some of my remarks are critical, I should stress that they intend no disrespect, nor can they lessen the lasting merit of Jambuvijaya's work. Two manuscripts form the basis for Jambuvijaya's edition; a Sarada manuscript in the Oriental Institute, Baroda, and a manuscript in Jaina Devanagari script, at that time in the possession of the well-known Jaina scholar Muni Punyavijaya. No other manuscripts are mentioned, and we may' assume none were known to him. The Jaina Devanagari manuscript has now passed into the collection of the L.D. Institute, Ahmedabad.26 I am indebted to the kindness and efforts of Muni Jambuvijaya, Prof. A. Wezler and the authorities of this institute for a photocopy. The Baroda Sarada manuscript I was allowed to photograph. 251 may refer especially to the English introduction contributed by Anantalal Thakur and the review by E. Frauwallner in the WZKSO, 1962. 26 As far as I am aware, it has yet to be catalogued. 12

Page #13
A collation of these two manuscripts with Jambuvijaya's edition revealed a very considerable number of cases where the manuscripts have been misread, wrongly reported or not reported at all. In several of these cases a misreading appears to have led to errors entering the edited text. I shall briefly illustrate this by an example which is particularly suitable because it can be checked by anyone who has access to the edition, simply by examining for himself the plates contained in it which reproduce the beginning of the commentary as it is found in the two manuscripts. According to Candrananda, the VS was taught by the sage Kanada to an unnamed brahmin who came to him with questions. The first word of the first sutra, atha, conveys according to the commentary the sense of anantaryam; immediately after he has been questioned as to the dharma, Kanada announces his intention to expound on dharma. The second word of the sutra is atah, and on this the commentary as edited by Jambuvijaya reads 'atah 'sabdo 'pi vairagyaprajnakathaparipakadikam sisyagunasampadam hetutvenapadisati, yasmad ayam sisyo gunasampada yuktas tato 'smai prasnebhyo 'nantaram dharmam vyakhyasyamah. The only variant reading given by the editor is P (the siglum for the Jaina Devanagari manuscript) 'nopadio for onapadio. Now it is not clear to me what kathaparipaka as a virtue of a student would mean, and I should think that any reader would consider the possibility of textual corruption here. So let us have a look at the manuscripts as reproduced in the edition. First the Sarada manuscript, the opening leaf of which is to be found as Plate I. I transcribe the manuscript's reading of this sentence, starting in the middle of line 9, introducing word-division but making no other alterations to the text. atahsabdo pi vairagyaprajnakasayaparipukadikam sisyagatasampadam hetutvenapadisati27 yasmad ayam sisyo gumasampada yuktah tato smai prasnebhyo nantaram dharmam vyakhyasyaman. This is rather alarming; in a single sentence we find two substantive differences from the edition, neither of which is reported in the critical apparatus. One of these, sisyagatasampadam for sisyagunasampadam, may be rejected as an error, particularly in view of the subsequent gunasampada yuktah. The other, however, provides us with the solution for our difficulty, for the reading kasayaparipakadikam for kathaparipakadikam yields good sense and is indeed immediately convincing. But let us see what Jambuvijaya's other manuscript reads here before 27 After this there is a small mark which should probably be interpreted as a half-danda. 13

Page #14
drawing conclusions. The text of the Jaina Devanagari manuscript may be transcribed as follows from Plate V, beginning in line 5. atahsabdo pi / atahsabdo pi vairagyaprajnakatha(ya) paripakadikam sisyagunasampadam hetutvenapadisati / yasmad ayam sisyo gunasampada yuktas tato smai prasnebhyo nantaram dharmmam vyakhyasyamah. At the point we are presently concerned with the scribe wrote kathayao, but this was later altered whether by the scribe himself or another we cannot tell--to katha, by the cancellation of the ya. Clearly the scribe had misread sa for tha (an easy mistake to make, especially from an examplar in Sarada script) 28 and the reading katha is a wrong correction of the senseless kathayao. It may therefore be regarded as certain that "kasayao is the correct reading, and from his acceptance of kathao, with not so much as a note in the apparatus, we are forced to conclude that in this case at least the editor has been less than scrupulously careful in transcribing his manuscripts and in critically reading his own text. We may note that another substantive variant of the Jaina Devanagari manuscript has not been reported in Jambuvijaya's apparatus, though this is admittedly only the clear dittography of atahsabdo pi. On the other hand the single variant which is given in the apparatus is a false one, for the manuscript clearly reads hetutvenapadisati as transcribed above, and not hetutvenopadisati as the apparatus suggests. Despite the fact that this is no isolated example, I should repeat here 28 There are other places too where the Jaina Devanagari manuscript shows traces of having been copied from a Sarada examplar. For instance, in a few cases jihvamuliya before k has been misread by the scribe as tk; thus in 1.1.28 the section of the manuscript which gives the sutrapatha separately reads samyogavibhagat karmanam for samyogavibhagah karmanam. Another case which should be noted is the sutra 8.10. The edition reads this dravyesv anitaretarakaranat karanayaugapadyat. No variants on this are given in the critical apparatus, but in the urddhipatrakam Jambuvijaya reports that the Sarada manuscript and the section of the Jaina Devanagari manuscript which gives the sutras within the commentary read thus, while the first part of the Jaina manuscript, giving the sutrapatha alone, reads dravyesv itaretarakaranat karanayaugapadyat. He then adds that 'dravyesv anitaretarakaranah karanayaugapadyat' iti patho 'tra samicino bhati (p. 231). He certainly is right about this, but two points need to be remarked on. First of all, the Sarada manuscript is in fact not available here as a witness. As was correctly noted in the last entry in the apparatus on p. 62, a large section, including the text of 8.6-13, has been left out in the Sarada manuscript (and this applies also to the other Sarada manuscript, not known to Jambuvijaya, which is introduced below). So it is the Jaina Devanagari manuscript alone which is present here. Secondly, the reading karanat which we find in both sections of the manuscript, can with virtual certainty be explained as misreadings of (Sarada) jihvamuliya. The fact, then, that even the Jaina Devanagari manuscript most probably descended from a Sarada manuscript, is an additional piece of evidence tending to suggest that Candrananda was a Kasmira.

Page #15
that Jambuvijaya's edition is an impressive achievement. Nonetheless, reexamination of the manuscripts does frequently bring to light readings which were either overlooked or wrongly reported in his edition. Regardless of whether or not the text of a new edition were to differ in many places from Jambuvijaya's edition, it would be sufficiently justified, I feel, if it succeeded in reporting the manuscript evidence more accurately, and thus allowed the user of it to judge the authority of the text for himself. Furthermore, I am happy to say that the manuscript basis for a new edition can now be extended somewhat further. In the Bhandarkar Oriental Research Institute, Poona, there are in fact two manuscripts of Candrananda's commentary which apparently have hitherto escaped notice. One of these, No. 403 of 1875-76, is a manuscript in Sarada script, while the other, No. 99 of 1873- 74, is in Jaina Devanagari script and is dated samvat 1931 (A.D. 1874). The latter proves to be of very little significance, for as I hope to demonstrate in detail elsewhere,29 it is virtually certain that it is an apograph of the manuscript in Ahmedabad. The Sarada manuscript, however, seems to be a new witness for the text. It is closely related to the Sarada manuscript used by Jambuvijaya, sharing quite a number of common errors, but each has errors and omissions of its own which rule out the possibility that either is an ancestor of the other. Instead, the evidence strongly suggests that both are descendants (I suspect even direct apographs) of a single hyparchetype; a manuscript which is lost or at least has not yet been brought to light. On the basis of all the manuscript evidence, conclusions differing from those of Jambuvijaya are sometimes possible not only in the text of the commentary but also as to the reading of some sutras. A single example. In 2.2.26 the reading accepted by Jambuvijaya is adityasamyogad bhutapurvad bhavisyato bhutac ca praci. In the critical apparatus he notes that O the Sarada manuscript in Baroda) reads samprayogad instead of samyogad. This is correct, and I may add that the other Sarada manuscript, in Poona, reads the same. What Jambuvijaya has however failed to record, either in the critical apparatus on the page or in the urddhipatrakam, which contains additional variants for the text of the sutras, is that the portion of the Jaina Devanagari manuscript which gives the sutrapatha separately also reads samprayogad. It is therefore only in the sutra as found within the 2 In the introduction to my forthcoming edition of Candrananda's commentary. 30 This vrddhipatrakam is introduced by Jambuvijaya with the words asmin granthe O. P. PS. madhye ye sutrapathabhedas te tatra tatra tippanesupadarsitah | tathapy asmadana. vadhanad ye 'vasistah pramarjaniya va pathabhedas te 'tropadarsyante katipayanam sutranam granthantaresuddhrtena sutrapathena saha tulana catropadarsayis yate. (p. 227). 15

Page #16
commentary in the later portion of the Jaina Devanagari manuscript that the reading accepted in the text is to be found. Since samprayogad, on the other hand, is attested in both the streams of transmission, it must be accepted as the reading most probably followed by Candrananda. In the commentary on the same sutra, the word reoccurs; Jambuvijaya again reads adityasamyogad, with the Jaina Devanagari manuscript, and this time fails to report that his Sarada manuscript once more has adityasamprayogat, a reading which, again, is shared with the Poona Sarada manuscript. The very fact that the most wide-spread recension of the text, that of SM, reads adityasamyogad in the sutra, renders it at least marginally more plausible that the Sarada manuscripts preserve the 'original' reading. It is interesting to note that adityasamprayogad is also supported by the manuscript in Malayalam script described in the previous section, T (see the third example quoted in section III above). This reading thus does not seem to be a purely local, Kashmiri one. Finally I should mention that there is a possibility that still other manuscripts of Candrananda may survive. Only recently I learned of the existence of a Sarada manuscript of a Vaisesikasutravstti in Ujjain and a Devanagari manuscript said to bear the same title in Jammu.32 There is more than a slight chance that one or both of these manuscripts may turn out to contain the text of Candrananda's commentary. I hope to have an opportunity to examine these manuscripts in the near future. I turn now to the next oldest extant commentary on the VS, that by Bhatta Vadindra. The situation with regard to the commentary by this scholar is somewhat complicated (as may be witnessed by the fact that even some very recent publications seem to have fallen victim to a certain confusion) and the scope for textual improvement here is considerably greater than with Candrananda's commentary, as I hope to be able to show. In 1957 a slin volume appeared containing the text of the VS together with what the title called an anonymous commentary.33 As the editor, A. Thakur, informs us in the introduction, the text was based on a Deva 31 In general, I think it is not unfair to say that Jambuvijaya is less careful in reading and reporting his Sarada manuscript, O, than he is with his Jaina manuscript, and that he also seems slightly biased at times in favour of the readings of the latter. 3?I am indebted for this information to Mr. Dominic Goodall, Wolfson College, Oxford. 33 For the details of this publication see under V in the first section of the bibliography below. 16

Page #17
nagari transcript from a single palm-leaf manuscript in Malayalam script.34 This manuscript had been mentioned five years earlier by V. Venkatarama Sharma, in a very brief article published in the Journal of the Oriental Insitute, M.S. University of Baroda.35 The transcript, riddled with errors and lacunae, was sadly defective as a basis for the constitution of a reliable text, but on the other hand the importance of the work was so great-remember that at this time Candrananda's commentary had not been published that we may be grateful indeed to Thakur for undertaking the task of its editor. 36 Although the text was published as the work of an anonymous author whose date could not be precisely fixed, in the introduction Thakur stated that 'it agrees with the sutra tradition followed by Bhatta Vadindra of the South. A preliminary study of the available portions of the Kanadasutranibandha of Vadindra has convinced us that the present commentary is an abridged version of this onibandha.937 Some time later, after a more thorough comparison of the text he had edited with manuscripts of the commentary by Bhatta Vadindra on the VS (BV),38 Thakur concluded 34 Note that some information, such as the fact that the original manuscript was a palmleaf one and that the transcript used was into Devanagari, is to be found in the Sanskrit bhumika (in this case on p. 23), but is omitted in the corresponding portion of the English introduction. There are numerous other differences of content and wording as well. 3* Sharma 1951, 226-227. The wording used by Sharma, '(recently I was able to procure a palinleaf manuscript containing an unknown commentary (urtti) on the Vaisesika-sutras, with the text,' suggests that the manuscript was actually owned by him at the time. From the introduction of Thakur's edition, however, one gathers that the manuscript had been in the possession of V.A. Ramaswami Shastri (who had however passed away by the time the introduction was written); '...a transcript of a single Malayalam manuscript prepared and supplied to us by the late lamented scholar, V.A. Ramaswami Sastrin' (p. 7), ... matskeyarn ... vio eo ramasvamisastrimahodayasyantika asit (p. 23). 36 As Thakur himself elegantly puts it, atra trutibahulyam asman sthagayati sma visayagauravam ca prakasanavidhau prerayali smety ubhayata akrsyamanair asmabhih prakasanam evorarikrtam (p. 23). *English introduction to the edition of V, p. 8. The corresponding passage in the Sanskrit bhumika reads Trayodasasatakasthitasya sarikarakirkaraparanamno bhattavadindrasya kanadasutranibandhena prastutasya granthasya drdhah sambandho viharigamadrsa avalokito 'smabhih | iyam hi vyakhya visayasamyena bhasasamyena ca tasya nibandhasyaiva sarasaingraharupa ily abhati (p. 26). It is something of a problem to determine what we should call this text. The name Vaisesikasutravarttika is found in three of the four colophons quoted from the manuscripts by Thakur (1960, 23 and 26); the fourth uses the name Kanadasutravarttika. These colophons are again reproduced in the printed text. The colophon of the section commenting on the first three sutras, attributing it to Bhatta Vadindra's patron, the Yadava king Srikysna, reads as follows: iti sriyadukulakamalakalikavikasabhaskarabhupalalalitamaharajadhirajasriktsnabhupalaviracite tarkasagaranamni vaisesikasutravarttike 17

Page #18
that the former was indeed 'nothing but an abridged version of the Nibandha giving mainly the interpretation of the sutras. It must have been prepared by Vadindra himself or some of his followers for those who were interested in the purport of the sutras and had no aptitude to enter into the abundant discussions of the Nibandha.'39 The publication of BhV itself, edited again by Thakur, unfortunately only followed very much later, in 1985. This edition is in many respects less satisfactory than the edition of V; we are, for instance, given no information on the manuscript basis of the text.40 Since variant readings are never given, one suspects that the text may be no more than a transcript, sporadically corrected, of a single manuscript.41 This publication also contained trisutrivyakhya samapta (Thakur 1960, 23 n.2; BhV 57). This provides us with another title, and one which is explicitly said to be a real name rather than a description or generic name (as Vaisesikasutravart tika can be taken to be). Of course the extent to which colophons should be relied on in these matters is debated. None the less, this particular colophon is probably authorial rather than scribal, bearing in mind the fact that it is not a concluding colophon of the work or an adhyaya or ahnika, but separates what is supposed to be the work of the royal patron (we may agree with Thakur in taking this to be a polite fiction) from that of Bhatta Vadindra himself. Certainly the style of this colophon is more flowery than one would expect a scribal one to be. And it seems more likely that a complimentary colophon should have been composed by Bhatta Vadindra than by a later scribe unconnected with the court of Srikrsna. For these reasons, I am personally inclined to believe that the title Tarkasagara may be the one bestowed on the work by its author, and that he also calls the work a Vaisesikasutravarttika. The titlepage of the edition refers to the text as Vaisesikavarttika, and this form of the title is also used by Halbfass (e.g. Halbfass 1992, 79). As far as I am aware, there is no basis for this title in the colophons or the work itself. Most likely it is an abbreviation of Vaisesikasutravart tika introduced by Thakur or perhaps even more probably-by the publishers. In his introductions to the edition of V, Thakur had called the text Kanadasutranibandha, as we saw above; he also uses this form in the introduction he contributed to Jambuvijaya's edition of the VS together with Candrananda's commentary. Finally, Thakur's 1960 article refers to the work as Kanadasutranibandha, on the basis of the second half of the opening verse of the commentary: kanadasutrasya maya nibandho vidhiyate sarikarakinkarena. The reprint of V as an appendix to Thakur's edition of BhV is given the name Nibandhasara, an allusion to the last two possible titles of BhV. But here, again, there is no manuscript authority for such an appellation of V, and this is probably to be regarded as a title made up by the editor or the publisher. 39 Thakur 1960, 27. Thakur reaffirmed his opinion that the brief 'anonymous' commentary was an abridgement of Bhatta Vadindra's voluminous one in the introduction he contributed to Muni Jambuvijaya's edition of the VS with Candrananda's commentary (p. 17). 40 There is no introduction by the editor, though we find an amukham by Dr. Jayamanta Misra and an aumakramikam by Ananda Jha. Neither of these provides the sort of information that a student of the text looks for first. 41In his article on this text Thakur had mentioned that three manuscripts in Malay 18

Page #19
a reprint of the text of V in an appendix. This has however practically no value; it introduces new misprints, contains no improvements (though a number would have been possible on the basis of BhV; cf. below), and does not even incorporate the corrections contained in the list of addenda and corrigenda appended to the original edition of V. Finally, a second appendix contained another welcome editio princeps, this time of the ninth adhyaya of the anonymous commentary on the VS written at the Sena court (S)-yet another text on which Thakur had given valuable information in an earlier article. 42 The commentary on the tenth adhyaya, surviving, like that on the ninth, in a single manuscript in the Asiatic Society, Calcutta, remains unpublished. Whatever its shortcomings, this publication allowed scholars with no direct access to the manuscript material to compare BhV and V for themselves, at least for a sizeable portion of the text. And in my opinion, Thakur's judgement is most probably correct. The difference in length between the two commentaries is very great indeed. The available portion of BhV covers 256 pages of Thakur's edition; the corresponding text of V, as reprinted in the appendix of the same edition, merely 26. But almost each sentence of V can be found also in BhV, though sometimes with slightly different wording. And in numerous places the published text of BhV allows us to correct what are clearly errors in V. A small example.3 On p. 3, line 22-23 (p. 156 line 15-16 in the reprint in BhV, appendix 1), we read in V dviprthag ityadivyavaharasya dvitvadyavacchinnaprthaktvad evopapatter iti kecit. The corresponding passage in BhV is to be found on p. 67, line 5-6, where we find dvau prthag ityadivyavaalam script are preserved in the Madras Government Oriental Manuscripts Library. His quotations are all based on transcripts of these manuscripts in the Mithila Institute. I doubt that Thakur used the Malayalam manuscripts themselves. From his description of the extent of the manuscripts, it appears that for a large portion of the preserved text at least two manuscripts should be available. The complete lack of variant readings in the edition is therefore odd. My own guess is that the edition is basically nothing else than a transcript of the largest of the Mithila Institute transcripts, and has not been collated against the other two transcripts. 42 Thakur 1965. This article contains the basic information on the manuscript material which one would have looked for in an introduction to the edition. 43 More significant examples could be given, but would require very much more space to set forth and discuss. Let me just mention here, without a detailed demonstration, that the text of the sutra numbered 2.1.12 in V (corresponding to C's 2.1.13), as well as the commentary thereon, should be emended in the light of BhV. The reading of the sutra should be adravyatuena nityatvam uktam. Some other cases where BhV confirms a correction made on the basis of the palm-leaf manuscript of V will be given below. 19

Page #20
harasya dvitvavacchinnapsthaktvad evopapatter iti kecit. Since we are here speaking of vyavahara, the reading dvau prthag is clearly to be preferred. And in fact this is exactly what the palm-leaf manuscript (see below) of V reads, so that the reading dvipsthag in this case probably originated as an error in the Devanagari transcript which was Thakur's sole source for V. . Thakur's judgement of the relationship between these two texts is therefore most probably to be accepted. Though we should certainly remain aware of some problems and difficulties, his characterization of V as an abridgement of BhV,44 retaining especially the portions of direct relevance to the interpretation of the sutras and omitting many lengthy discussions and digressions, is clearly more accurate and helpful than the mere statement that both works are indebted to Udayana and use similar versions of the Sutra text.145 "It might be objected that Bh V could equally well be an expanded version of V. This possibility cannot perhaps be completely ruled out, but the probabilities are weighted very heavily against this in my opinion. It seems unlikely that Bhatta Vadindra should omit to compose one or more opening verses for a commentary on the VS, even a brief one. Nor does it seem plausible to me that he should have made the commentary of another author the basis for his own fuller one, following it so faithfully as to hardly omit a word in it, and yet fail to acknowledge the fact. Bhatta Vadindra is I think too much an original scholar and indeed idiosyncratic thinker for that. In addition, I suspect that there is internal evidence which points to V indeed being an abridgement made on the basis of BhV. I must however postpone discussing this point, which is obviously complicated by the fact that both texts are only available to us in mutilated and sometimes corrupted forms. Another question which is more difficult to settle is whether Bhatta Vadindra is himself responsible for abridging Bh V into V or whether this is the work of another hand. And in the latter case, is the abridgement none the less roughly contemporaneous with the composition of BhV-is it for instance an extract made by a student of Bhatta Vadindra for his own use or is it a much) later recast? This question is of importance for our evaluation of the occasional sentences in V which do not seem to have a counterpart in BhV. Once more, I can not present evidence in full-an attempt to settle the matter would require very close study of the two texts together with the other extant works of Bhatta Vadindra, and would in effect almost have to be preceded by re-editing both versions-but I personally feel that it is quite unlikely that the abridgement is authorial. 45 Halbfass 1992, 84 n.25. Also in the other passages of this important book where Halbfass refers to Bhatta Vadindra's commentary, it appears that he regards BhV and V as completely different texts. Thus on p. 75, he speaks of 'several apparently older commentaries; that is Candrananda's Vitti, Bhattavadindra's Varttika, and the anonymous commentary edited by Anantalal Thakur,' and on p. 79 he calls V 'an anonymous Vyakhya, which may be several centuries older than the Upaskara.' Nowhere does Halbfass give his reasons, if there are any, for differing from Thakur's judgement. I do not however wish to imply that there are no discrepancies at all between the two commentaries. But I suspect that most of the cases where they seem to differ in substance (as opposed to mere variation, usually slight, in wording) are to be explained as resulting 20

Page #21
The text of V as printed is in many respects problematic and unsatisfactory. For this one can hardly reproach Thakur, for the material he had to work with simply was too poor and scanty to establish a reliable text. On the basis of Candrananda's text and commentary some improvements were possible, particularly in regard to the sutra text followed by Bhatta Vadindra, and both Muni Jambuvijaya (in the second appendix of his edition of the VS with C) and M. Nozawa (in an article which appeared in 46 Further 1974) put forward a number of emendations to the sutrapatha." correction of the text of V, at least for the first two ahnikas, became possible with the publication of BhV, as has been remarked above. But even so, further improvements are rather badly needed, especially for the portions where the corresponding text of BhV is not available. Fortunately there is a source which will allow an advance in the right direction. This is nothing else than the palm-leaf manuscript of V, from which the transcript used by Thakur was made. This manuscript was acquired rather recently by the Kerala University Manuscripts Library, where it bears the number 21600C.47 The route by which it came into the library's collection is not completely clear. According to the library's records, its last owner was K.V. Sharma. There can however be little doubt that this is indeed the very same manuscript which was described by V. Venkatarama Sharma and transcribed for Thakur; for that the manuscript agrees too closely with the edition. For instance, the lacunae in the edition which Thakur usually attempts to fill up by conjecture nearly always correspond to the places where text has been lost due to the margins of the palm-leaf manuscript being broken. The condition of the manuscript seems to have deteriorated only slightly from the time that the transcript Thakur used was made. The margins of from the defective state in which both have reached us. Indeed, as will be shown directly below, consultation of the original palm-leaf manuscript of V frequently allows restoration of a text substantially closer to that of BhV. Another factor to be taken into consideration is the possibility that errors were made during the process of abridgement; for instance, in some occasions, the sense of a passage may have been altered, even perhaps against the intention of the abridger, by the omission of certain sentences or words. Of course this is only likely if the person responsible for the abridgement was, as I suspect, different from Bhatta Vadindra himself. This type of change or error can however probably not be identified with certainty because it can never be excluded (and is usually more plausible) that omissions of the kind I am thinking of are to be put down to scribal error. 46 The two scholars do not however always agree in their emendations. 47 This manuscript too is not listed in Bhaskaran 1984. I have consulted it from photographs. 21

Page #22
most of the leaves are damaged, usually resulting in the loss of a few aksaras, but in some case rather more than that. Only in a few places does it appear that syllables which were in Thakur's transcript, and hence presumably legible in the palm-leaf manuscript when the transcript was made, have now been lost, due to further crumbling of the margins. Several folios have been considerably darkened, most probably due to the effects of smoke, but this has not led to text becoming illegible. The hand is early Malayalam, perhaps of the seventeenth century. 48 In addition to the commentary we are concerned with, it contains the Nyayadipavals and a commentary thereon. Bhatta Vadindra's commentary covers folios 110-147. At the end, some stray folios occur containing part of the end of Saktibhadra's well-known play, the Ascaryacudamani. As was to be expected, a comparison of the palm-leaf manuscript with the printed text brought to light a substantial number of cases where corrections are possible. The transcript undoubtedly contained a number of misreadings, and also has on occasion omitted passages, usually due to homoeoteleuton or homoeoarcton. Furthermore, where the original was damaged, the transcript probably did not indicate the number of syllables which may have been lost, so that some of Thakur's conjectures are less plausible simply in view of the space they would have taken up. Given that Thakur was unable to make use of the original manuscript itself, this sort of problem was of course well-nigh inevitable. One helpful feature of the palm-leaf manuscript is that the sutras are usually set off from the commentary by the addition of tiny dots at their beginnings and ends. These are the only punctuation marks found in the manuscript. Thakur's statement that 'the manuscript does not distinguish the sutras from the commentary49 thus applies only to the transcript, and demonstrates the fact that he never saw the original manuscript. But even with the palm-leaf original at our disposal, to establish a satisfactory text is a formidable task-in several cases an impossible one. The manuscript contains a rather large number of scribal errors, and numerous passages are viciously corrupt. Larger lacunae can certainly never be restored with anything approaching certainty, unless another manuscript should come to light. The character and style of the text also does not make Such would be my guess, and in this I find myself in agreement with Sharma's estimate of the age of the manuscript as some three hundred years (cf. Sharma 1951, 226). "From the English introduction, p. 9. The sentence has no parallel in the Sanskrit bhumika.

Page #23
matters any simpler. Bhatta Vadindra's style is often intricate, 50 and his thought, influenced by Udayana, sophisticated. His explanations of sutras are often surprising, not to say unnatural. Still, with care, it should be possible to arrive at superior readings in a great many cases. Here I shall confine myself to giving a few examples which are relatively clear and should not require too much discussion.51 More extensive lists of corrections and emendations are intended to be published elsewhere. First a few cases where text has inadvertently dropped out in the printed edition. The reading of the edition is given first. References are to the page number and line of Thakur's 1957 edition (rather than to the reprint in the appendix of his edition of BhV). The portions between square brackets in the quotes from the edition are Thakur's proposals for filling up real or conjectured lacunae. 1 In the commentary on VS 1.1.3 (tadvacanad amnayasya pramanyam), amnayasya is explained by Bhatta Vadindra by adding srutismrtitihasadeh (2.17). Thus the edition; the manuscript has srutismrtitihasapuranadeh, the same reading as is found in BhV (p. 13 line 2). 2 In the long (even in the abridged version!) and intricate commentary on the sutra kriyavad gunavat samavayikaranam iti dravyalaksanam (C's 1.1.14, numbered 1.1.15 in the edition of V and 1.1.14 in the edition of BhV52) we find a sentence which reads as follows. napi vrddhavyavaharad eva dravyasabdavacyatvasiddhe[r anumanasya vaiyarthyam, ekasadhanenanyasadhanasya, anyatha anumanena dravyasabdavacyatvaprasiddhe]r vrddhavyavaharavaiyarthyasya durvaratva! (8.7-9). Thakur deserves full credit here for realizing that the text available to him was corrupt and for correctly diagnosing the location and cause of the corruption: loss of text due to homoeoteleuton. The manuscript reads napi vrddhavyavaharad eva dravyasabdavacyatvasiddher vyatirekiyaivayyartthyam (read vyatirekivaiyarthyam) tasyanupajfvyatva; anyatha vyatirekenaiva dravyasabdavacyatvasiddher 50 Though at least the abridged version makes for easier reading than the long one. 51 For keeping discussion of the often considerable problems at a minimum here I must plead shortage of time and, above all, space. 52 Here there is a discrepancy between V and BhV that cannot be easily accounted for. But na tu karyabhavat karanabhavah which is numbered in V as 1.1.14 is no doubt not to be taken as a sutra which is meant to go here. It is a quote of 1.2.2. Perhaps text has been lost in BhV which contained this quote and the following passage in V which seems to have no equivalent. 23

Page #24
vrddhavyavaharavaiyartthyasya durvaratvat. This agrees exactly with the corresponding passage in BhV (p. 103 l. 19-22), except that the latter has duspariharatvat. 3 In the commentary on 6.2.1, drstanam drstaprayojananam drstabhave prayogo 'bhyudayaya, we read [evam sati] gobrahmanadyuddesena tyagah svargasddhanadharmasddhanam [vedapramanejna bodhyata iti (61.4-5). The manuscript has prayogo brahmanadyuddesana (read brahmanddyuddesena) tydgah tatha ca dratanam hiranyasamidajyacarupurodasadinam brahmanadyuddesena tyagas svargasadhanadharmasadhanataya... (2-3 aksaras lost here) na bodhyata iti. Again, the loss of text was clearly caused by eyeskip. 4 In the edition, the sutra adoso 'nupadha (C's 6.2.5) is not to be found; after the commentary on 6.2.4 (numbered 6.2.5 in V) the edition continues with the sutra [yad] istaruparasagandhasparsam proksitam abhyksitam ca tac chuci (61, 18; the sutra corresponds to C's 6.2.6). The missing sutra is however present in the manuscript. After the final word of the commentary on (C's) 6.2.4, the following should be added: adoso nupadha [6.2.6 C's 6.2.5] adusto bhisandhinupadheti (read 'bhisandhir anupadheti) laksanam. In this case an entire sutra and its (brief) commentary has dropped out due to homoeoteleuton; the commentary on 6.2.5 (C's 6.2.4) also ends on the word laksanam. In the following sutra, the [yad] was added by Thakur to make the sutra agree with SM; it may now be removed, for it is not supported by the commentary, and is not found in C's text. = Finally, some corrections of misreadings which do not involve omission of text. 5 In V's 3.1.3, corresponding to C's 3.1.2, the reading we find in the printed text is indriyarthaprasiddher indrigarthebhyo 'rthantarasya hetuh. The manuscript reads indriyartthaprasiddher indriyartthebhyo rtthantaratve hetuh. The aksara tve has been added under the line (but by the same hand) and is a little difficult to read, but I believe there is no real doubt possible about the reading. I propose emending to read with C indriyarthaprasiddhir indriyarthebhyo 'rthantaratve hetuh. The all too brief commentary-it merely runs prasiddhyasrayasyeti sesah seems to me to support prasiddhir rather than prasiddher. 24

Page #25
6 6.2.2 is read as follows in the edition: abhisecanopavasabrahmacaryagurukulavasavanaprasthayajnadanaproksanadinaksatramantrakalaniyamas cadrstaya. C and SM have vanaprasthya for vanaprasthadeg. The palm-leaf manuscript indeed reads as printed by Thakur. In the commentary on this sutra, however, we find the following remark: vanad vanam pratisthata iti vanaprasthah | sa tu trtiyasrami tasya karma vanaprastham (61.10-11). The manuscript has sa tu for na tu, and vanaspatyam for vanaprastham. I suggest that we should read and punctuate vanad vanam pratisthata iti vanaprasthah | na tu trtiyasrami tasya karma vanaprasthyam. Bhatta Vadindra's intention is, I believe, to explain that the neuter noun vanaprasthya is derived from the masculine noun vanaprastha by addition of the taddhita suffix SyaN in the sense of the activity or occupation of a person (karma), in accordance with Pan. 5.1.123. And the masculine noun vanaprastha is to be understood as meaning 'one who goes from forest to forest,' i.e., presumably, a wandering ascetic, and not as someone in the third stage of life (as the word would ordinarily be taken), who would be-as Candrananda says-one who leaves from his house to the forest. Compare Candrananda's commentary ad loc.: sastravidhina gehan nihsrtyaranyam prasthito vanaprasthah, tasya karma vanaprasthyam (C p. 48 1. 13-14). If I am correct, we should then also emend the sutra to read vanaprasthyadeg with C and SM. It should be noted though that A and T both have 'vanaprastha". VI The previous sections have done little more than present some notes on the manuscript tradition of the VS and the commentaries by Candrananda and Bhatta Vadindra. A more thorough treatment would require very much more time and space than is at my disposal just now. Nonetheless I hope that some of the readings discussed above, and the corrections of printed texts proposed on the basis of manuscript readings, may prove of interest to fellow students of the Vaisesika. To conclude, I should like to venture onto what is in a sense even trickier ground, and offer a few general remarks on questions of method. My apology for the fact that most of my observations are obvious, not to say banal, can only be that I know from my own experience that such basic points, or their implications, can all too easily be forgotten. I hasten to add also that I am most painfully aware of how far the work presented above, which can at best be described as preparatory, 25

Page #26
falls short of the ideal which is broadly sketched below. Those who undertake to study classical Indian philosophy must inevitably base their researches in the first place on texts. And since it is practically speaking never the case that we possess the author's autograph manuscript, certified beyond doubt, and unambiguously legible,53 it appears to me to follow inevitably that textual criticism is an essential discipline.54 And especially in cases where the surviving manuscripts are all many centuries later than the texts they transmit-and this is the situation with all of our early texts-it would appear to be self-evident that it is our task to attempt to collect all available evidence, both primary and secondary, and to bring to bear all we can learn about the ways in which texts were transmitted and altered in the hope of being able thus to determine as far as possible what the original form of the text was and how it changed over time. I would like to stress that recovery of the original is, in my view at least, not necessarily the highest, and certainly not the only goal of the text critic. Rather, it is the reconstruction of the history of the text, which is essential for the recovery of the original, but which often includes far more. For it requires, one might say, that we enter into the mind and thought not only of the writer but also of all those who have influenced its transmission. It demands, in addition to the more mechanical and basic skills, sensitivity to historical development, awareness of why and how a text may have been changed-and this means an understanding of the text as a part of the culture of which it forms a part. Rather than making the reconstruction of a single moment of creation our goal, this approach attempts to grasp the development of the text in its entirety. Over and above the individual thinker, the critical study of texts can shed light on Indian culture as something changing and developing. To be a little more concrete; if we wish to reconstruct the original text of the VS-if one can profitably speak of such a thing-we will need to identify the accretions and changes to the text. In this we are faced by different kinds of problems. We shall have to determine what readings (including omissions and additions) may be purely scribal, arising from unconscious 53 In fact, the problems and disagreements of editors of modern English and American authors should warn us that even in the case of works available in autograph manuscripts or typescripts, it is by no means always possible to arrive at agreement on the precise text to be adopted. 54 Even scholars who work exclusively from printed texts can certainly benefit from studying the transmission of the texts they deal with. For instance, knowing which scripts the manuscripts of a text were written in, together with a grounding in palaeography and codicology, can clearly help in alerting one to corruption and dealing with it. 26

Page #27
a number of our sources for early Vaisesika are preserved in single, unique manuscripts. My plea is therefore in the first place that we should not forget how our knowledge ultimately rests on highly perishable documents, most of which have yet to be studied thoroughly. This is something which some who work exclusively with printed texts may occasionally lose sight of. This does not mean that I advocate all of us immediately leaving our desks to go in search of manuscripts, although I do think that such work should be kept up by a few at least. But we should remember not to accord the editions we have more authority than they deserve. Chance has played too great a role in determining which texts are now available to us as printed books, and in what form-the chance of one work surviving while another was lost; the chance of one being transmitted faithfully while another is corrupted by poor scribes or changed deliberately to suit the needs or taste of a later period; the chance of one being discovered while another molders in an unsearched stack of manuscripts; the chance of one finding a competent and sensitive editor while another suffers from the rough hands of an impatient scholar, all too quick to emend what he does not understand.97 With this in our minds, we would do well to be humble about the reconstructions we may arrive at of the thought of writers separated from us by so many centuries and the work of so many scribes. 57 A conservative editor, slow to admit that the text of his manuscripts is corrupt and loath to emend it, is likely to do less damage. 28

Page #28
BIBLIOGRAPHY I. Editions of the VS with commentaries C Candrananda's commentary on the VS. Vaisesikasutra of Kanada, with the commentary of Candrananda. Crit. ed. Muni Sri Jambuvijayaji. Baroda 1961. Gaekwad's Oriental Series 136. [A reprint of this edition appeared in 1982, but if at all possible, references should be made to the first edition. The reprint appears at first sight to be photo-mechanical but in fact contains misprints which were not in the first edition. The suddhipatrakam of the first edition has also not been reproduced, although many of its corrections have not been incorporated.] BhV Bhatta Vadindra's commentary on the VS, long version. Bhattavadindraracita-vaisesikavartika-krsnabhupalaracita-trisutriprakasajnatakarttrkavrttibhir vilasitam maharsikanada-pranitam Vaisesika-darsanam. Ed. Anantalal Thakur. Darbhanga 1985. Maharajadhiraja-kamesvarasimha-granthamala puspam 21. V Bhatta Vadindra's commentary on the VS, abridged version. Vaisesikadarsana of Kanada, with an anonymous commentary. Edited by Anantalal Thakur. Darbhanga 1957. [Reprinted as appendix one of the edition of BhV listed above. The reprint adds new printing mistakes and does not incorporate the corrigenda appended to the original edition.] SM Sankara Misra's commentary (Upaskara) on the VS. The Vaiseshika Darsana, with the commentaries of Sankara Misra and Jayanarayana Panchanana. Ed. Pandita Jayanarayana Tarka Panchanana. Calcutta 1861. Bibliotheca Indica, New Series XXIV. S 'Sena Court' commentary, by an anonymous author, on the VS. For an edition of adhyaya 9, see the second appendix in Thakur's edition of BhV. The tenth adhyaya survives in a palm-leaf manuscript in Newari script in the Asiatic Society, Calcutta. The sutrapatha of this commentary for both the ninth and the tenth adhyaya is given in Thakur 1965. 29

Page #29
II. Secondary literature Bhaskaran, T. 1984 Alphabetical Index of Sanskrit Manuscripts in the Oriental Research Institute and Manuscripts Library, Trivandrum, Vol. III (Ya to Sa). Edited and published by T. Bhaskaran. Trivandrum 1984. Frauwallner, Erich 1984 Der ursprungliche Anfang der Vaisesika-Sutren. in: E. Frauwallner: Nachgelassene Werke I. Auftrage, Beitrage, Skizzen. Herausgegeben von Ernst Steinkellner. Wien 1984. Osterreichische Akademie der Wissenschaften, Philosophisch-Historische Klasse, Sitzingsberichte, 438. Band. Veroffentlichungen der Kommission fur Sprachen und Kulturen Sudasiens, Heft 19. p. 35-41. Halbfass, Wilhelm 1992 On Being and What There Is. Classical Vaisesika and the History of Indian Ontology. Albany, N.Y. 1992. Kaviraj, Gopinath 1929 Some variants in the readings of the Vaisesika sutras. in: Princess of Wales Saraswati Bhavana Studies 7 (1929), 71-76. Nozawa, Masanobu 1974 The sutrapatha of the Vaicesikasutra vyakhya. in: Journal of Indian and Buddhist Studies (Indogaku Bukkyogaku Kenkyu) 23.1 (Dec. 1974), 474(24)-471(27). 1985 A comparative table of the Vaisesikasutra. in: Numazu Kogyo Koto Senmon Gakko Kenkyu Hokoku 20 (1985), 75-93. Sharma, V. Venkatarama 1951 Vaisesikasutras. in: Journal of the Oriental Institute, M.S. University of Baroda 1 (1951), 225-227. Thakur, Anantalal 1963a Textual problems of the Vaisesikasutras. in: Journal of the Bihar Research Society 49 (1963), 186-188. 30

Page #30
1963b Vatsyayana and the Vaisesika system. in: Vishveshvaranand In dological Journal 1 (1963), 78-86. 1965 Studies in a fragmentary Vaisesikasutravitti. in: Journal of the Oriental Institute, M.S. University of Baroda 14 (1965), 330-335. Wezler, Albrecht 1982 Remarks on the definition of 'yoga' in the Vaisesikasutra. in: Indological and Buddhist studies. Volume in honour of Professor J.W. de Jong on his sixtieth birthday. Ed. by L.A. Hercus et. al. Canberra 1982. p. 643-686. 31